W3CJSON-LD (JSON for Linking Data) bridges the gap between the simplicity of JSON and the richness of Linked Data. As a W3C Recommendation published 16 July 2020, it provides a lightweight syntax for serializing Linked Data in JSON, designed to easily integrate into deployed systems already using JSON. JSON-LD has become the dominant serialization for structured data on the Web, underpinning Schema.org markup, Verifiable Credentials, and the Activity Streams protocol.
Background
The JSON-LD effort began around 2010 within the W3C community, driven by the need to make Linked Data accessible to mainstream web developers. Earlier RDF serializations like RDF/XML and Turtle required specialized knowledge that limited adoption. Manu Sporny and Dave Longley (Digital Bazaar), Gregg Kellogg, Markus Lanthaler (Google), and Niklas Lindstrom led the version 1.0 design. JSON-LD 1.0 reached W3C Recommendation status on 16 January 2014. The JSON-LD Working Group, with editors Gregg Kellogg, Pierre-Antoine Champin (LIRIS, Universite de Lyon), and Dave Longley, subsequently developed version 1.1. JSON-LD 1.1 is a superset of 1.0, and documents created with the earlier version remain compatible.
Purpose & Scope
JSON-LD is primarily intended for use in Web-based programming environments, interoperable Web services, and JSON-based storage engines. It can be used directly as JSON without knowledge of RDF, or as an RDF serialization in conjunction with SPARQL and other Linked Data technologies.
Key capabilities introduced by JSON-LD include:
- A universal identifier mechanism for JSON objects via IRIs
- A context mechanism (
@context) that maps JSON keys to IRIs, disambiguating terms across documents - The ability to reference resources on different sites
- Language annotation for strings
- Datatype association for values such as dates and times
- Expression of one or more directed graphs in a single document
Set of Documents
JSON-LD 1.1 is specified across three companion W3C Recommendations:
| Document | Purpose |
|---|---|
| JSON-LD 1.1 | Core syntax and data model for serializing Linked Data in JSON |
| JSON-LD 1.1 Processing Algorithms and API | Standard library interface for expansion, compaction, flattening, and RDF conversion |
| JSON-LD 1.1 Framing | Deterministic reshaping of JSON-LD documents using frame templates |
Design Principles
JSON-LD is designed to not disturb already deployed systems running on JSON while providing a smooth upgrade path to Linked Data. Since the shape of JSON data varies widely, JSON-LD features mechanisms to reshape documents into a deterministic structure that simplifies processing. The @context mechanism allows any existing JSON to be given Linked Data semantics by providing a mapping from JSON keys to IRIs. The processing mode system ensures that 1.0 processors will not accidentally process 1.1 features.
Governance & Maintenance
JSON-LD is maintained by the World Wide Web Consortium. The JSON-LD Working Group derived its specification from the JSON-LD Community Group's Final Report. Development is coordinated via the GitHub repository at w3c/json-ld-syntax and the public-json-ld-wg@w3.org mailing list. An implementation report and test suite are maintained alongside the specification.
Notable Implementations
JSON-LD is the serialization format for Schema.org structured data used by Google, Bing, and other search engines for rich results. The ActivityPub protocol (powering Mastodon and the fediverse) uses JSON-LD for its Activity Streams 2.0 payloads. The W3C Verifiable Credentials and Decentralized Identifiers specifications rely on JSON-LD. Major libraries exist for JavaScript (jsonld.js), Python (PyLD, rdflib), Java (Apache Jena, titanium-json-ld), and other languages. The JSON-LD Playground provides an interactive environment for testing and visualizing transformations.
Related Standards
- RDF -- The abstract data model that JSON-LD serializes
- Schema.org -- The most widely deployed vocabulary using JSON-LD contexts